Remixing your art with AI

13468921

Comments

  • WendyLuvsCatzWendyLuvsCatz Posts: 38,211
    edited January 2023

    it's a lot easier to do stuff if your only stake in it is I have a computer, I want to have fun

    I can understand people who are using their skills to earn a living being pissed though which is why I have stated I would never use ai art commercially and don't think it should be for the reasons already discussed 

    but I also don't want people gatekeeping what I do on my own computer 

    what books I read

    what music I listen to etc

    this also applies to what programs I use and if it's not explicitly against some federal or international law it's my choice 

    if it means a lot of people on the internet will not like me because of it I won't lose as much sleep as I do about countries invading each other or people being persecuted because of their gender roles etc, it's simply not that big a deal to me in the scheme of things.

    if legislation is passed limiting what can be scraped I am perfectly fine with it and have already said I agree with that, but I also want to be part of the exciting experience pf enjoying and exploring new technology not the one in the corner hogging all the crayons and pouting 

    for me playing with ai is what I do for recreation instead of playing a game like Cyberpunk 2077 or Genshin Impact, it's not a big part of my life but a fun one

    as to sharing the results on facebook I guess those preaching never share internet memes created using popular images or share or do art involving characters from movies, comics or video games

    Post edited by WendyLuvsCatz on
  • Geminii23Geminii23 Posts: 1,327

    WendyLuvsCatz said:

    it's a lot easier to do stuff if your only stake in it is I have a computer, I want to have fun

    I can understand people who are using their skills to earn a living being pissed though which is why I have stated I would never use ai art commercially and don't think it should be for the reasons already discussed 

    but I also don't want people gatekeeping what I do on my own computer 

    what books I read

    what music I listen to etc

    this also applies to what programs I use and if it's not explicitly against some federal or international law it's my choice 

    if it means a lot of people on the internet will not like me because of it I won't lose as much sleep as I do about countries invading each other or people being persecuted because of their gender roles etc, it's simply not that big a deal to me in the scheme of things.

    if legislation is passed limiting what can be scraped I am perfectly fine with it and have already said I agree with that, but I also want to be part of the exciting experience pf enjoying and exploring new technology not the one in the corner hogging all the crayons and pouting 

    for me playing with ai is what I do for recreation instead of playing a game like Cyberpunk 2077 or Genshin Impact, it's not a big part of my life but a fun one

    as to sharing the results on facebook I guess those preaching never share internet memes created using popular images or share or do art involving characters from movies, comics or video games

    Well, there is a lot to unpack there. I don't think I am the one to try. The discussions I have routinely with friends and colleagues surrounding the use of AI have more to do with when innovation crosses into exploitation. When I hear about people winning art contests using AI that is a problem. This stuff to me is no different than athletes cheating by using steroids. Using AI to create something is not creating anything at all.  And on a long enough timeline, everyone will become obsolete. I can only imagine what will happen when CEOs are able to offload writing, music, etc to AI. They will increase their bottom line and tens of thousands will be out of work. And "learn to code" isn't the answer or a realistic solution. We will have a permanent underclass in society. Programs like DAZ will also become obsolete. The fact that AI is able to comb through and curate millions of images online created by DAZ users on numerous sites from all over the world, and create interpretations and modifications in record time, should surely concern the DAZ content creators themselves.  What would be their incentive to create new models if there is no audience to purchase them anymore?  

  • WendyLuvsCatzWendyLuvsCatz Posts: 38,211

    now this is sad but says more about Reddit than AI

  • outrider42outrider42 Posts: 3,679

    To be fair, we often use AI already in some form. Denoisers, upscalers, and face reconstruction are all features based on AI software. However these are not reconstructing entire images and turning them into something else. These can do unintentionally goofy things sometimes, but they do not take a motorcycle and turn it into a goat.

    So the line for being able to copyright a work with AI will fall somewhere between that and a text to image generation. It will likely vary country to country. But I can see a lot of countries voiding text to image copyrights.

    I am not against AI art. I simply want it to be properly regulated so it doesn't get out of control. It can be fun to play with.

    It can certainly have high hardware requirements, which will no doubt be a barrier of entry. Sure, you can use various online models, but you are capped at how many free pics you can make, and their size. You also have to deal with potential censorship as seemingly innocent phrases may be banned and possibly get you banned from the service.  However doing AI on your local machine requires a ton of VRAM if you want a larger pic. If you want to train your own fork of AI you basically need 24gb of VRAM, and only a few GPUs in existance have this much. So a number of AI features are more demanding than Daz Studio Iray can ever be. As much as people in the forums complain about Daz using a lot of VRAM, these AI generators can totally obliterate some of your computers. You can use a gpu with 6 or 8gb, but you will be restricted to small images, and training is not available to you.

    So in this respect, Daz still has plenty of advantages. I can only render a maximum of 2048 by 2048 with Stable Diffusion using my 3090, this a lot larger than online AI sites allow (many are capped at 512 or 1024 pixels). But I can render significantly larger images with Iray or other render engines.

  • WendyLuvsCatzWendyLuvsCatz Posts: 38,211
    edited January 2023

    reply to a deleted post

    mods should delete also as made no sense

    this whole thread a dumpster fire TBH

    it was supposed to be remixing art with ai which I mostly do

    99% of my Stable Diffusion generated images use a 3D render or photograph of mine pre or post, 1% just text and no other modifications 

    Post edited by WendyLuvsCatz on
  • generalgameplayinggeneralgameplaying Posts: 517
    edited January 2023

    frank0314 said:

    When will it get to the point where it starts creating false situations amongst real people

    Certainly, while rather in the hands of humans, that's already happening. At least in terms of tools for targeting and spreading variations of texts, how to create impact amongst groups of people, at least. That for the text form. 

    With ai-created everything, you (as a manipulator with a good bit of resources at hand) are served the means for perfect feedback-loop testing. You could experiment with people's reaction towards changes, evaluate emotion, language and behavior, and keep adjusting in real time. Paired with recognition technology that is. A machine-learning-based system could keep adjusting in realtime, with whatever consequences for the poor visitors of the random place in the desert. Of course the governmental side of this has all sorts of dystopic possibilities to it as well.

    Puns aside, they probably already have a good bit of data, from all the donators who view ads and whose behavior gets tracked within social media. So maybe they could relate aggression for instance to some kind of shown content, already now. Just to be able to create content in real time, which matches certain criteria, like "totally unsuspicious", means a new era of manipulation (attempts). "They" is meant fuzzy, more in terms of "someone with access and abilities could...".

    Edit 2: "Content that induces..." is meant rather like "whatever sequence of whatever, in terms of whatever... makes whomever...", in terms of gathering bits and pieces about personalia and characters and psyche, so they get an idea about how to influence different kind of people in different situations. This will also involve, what happens to them in life and so on, so it likely is by no means trivial and perhaps not even possible in general. However, for creating "impact", it may be enough to influence groups to an extent, in terms of statistically influencing them. E.g. a political party that tears itself apart, based on several or most members following some kind of influence, may not be possible to be rescued by a few "knowing people with brains".

    Edit 3: "Statistically influence" by no means has to be random nor random people in random situations. It's always easiest to get a grip on people. when they're in confined spaces, e.g. all are upset, or "meeting for...", "rallying for...", "profession X [in situation Y]...". So it's not the all-random mixed-bag something, but instead researching ways to access people, where you can. The intrinsically enemic nature of this should be apparent, at least for the context of manipulation of people with access to lots of data in real-time. Of course it may be "necessary" to research for defending against it then, though this does mean implications for the idea of freedom too (Prototype: Bad guy observes and plays the keys, "good guy" observes and plays the keys for the sake of defence. Why let anyone a) observe, b) play the keys?).

    (How to make a section collapsible?)

    Post edited by generalgameplaying on
  • generalgameplayinggeneralgameplaying Posts: 517
    edited January 2023

    Future?

    • DAZ3D? I'd say precise 3D never will be out, but there will be need to keep up with competitors, and the market for simple and quick drawings with tasty effects may go to some cloud service, at least medium term. DAZ3D may have many edges with potential for high impact improvements (faster setup, help with placing/posing/animating/export/import), as well as cooperating with or incorporating generative ai services. E.g. they could distribute protoype renders from multiple angles with each product, and either cooperate directly, or craft it towards cooperation with some generative ai system(s). Perhaps it'll be feasible to have a whole model for the whole shop (how frequently can you train it?), or there will be advancements in generative ai, such that you not only have "a language model", but also "a posing model" and "an animation model" and "a placement model", as well as the possiblity to confine to certain assets being used or excluded during runtime. "Paint stick-figure 2 like in the image here" paired with the "knowledge" of the DAZ applicaitions condensed into auxiliary tools/models/parts of the system, may make quick rendering possible in pretty precise ways. Making work with 3D more efficient and simply faster, may be doable without means of ai, to some extent, though fast mock up change-preview might be doable by render + modification-ai. - or think of post-work inpainting and effects, simply. Of course they'll compete with cloud services to some extent, specialized on such. Imagine in that scenario, the stance of asset-stores vs. a "free to feast on" internet, in terms of scraping art made with their products - this could be hard competition in the end. Some licenses already contain "not for training ai with" - and who knows if lawmaking pulls the plug on restrictions the other day.
    • The very general generator? Uncertainty principle might prevent the precise "can do everything" thing. However, one day there may be advanced systems with multiple "grammar-like" auxiliary models used at the same time, as envisioned in the DAZ3D section for one case, which does a little less on the "arbitrary magic" side, but e.g. a combination of several tools, will still be easy to use, just way more precise to control than current text-to-image generators. Short term and for free, or as a cloud service rolling up the internet - hard to imagine very short term. After all common sense is hard to do in ai.
    • More specialized tools? These certainly are evolving already (not necessarily generative ai), but moreso will further evolve for sure. The high quality and precise tools will stay somewhat costly likely, similar to what is now, but allow for even faster and still high quality production, likely. For the relative masses (rather), perhaps pretty specific tools evolve, in cooperation or attached to generative ai services, resulting in less perfect but pretty nifty tools like for storyboarding, creating and modifying images and entire movies. These may have the greatest impact on the majority of (fast/simpl-ish) content producers. Impact could also hit movie/anime production to some extent, because you can make a movie for the favorite story of your child "in no time", with the effect of more children growing up, getting used and attached to the specific services. Of course many of nowadays producers will then use the better-quality tools, with uncertain continuation. "Good enough" may sweep the market at some point.
    • Race to the bottom? Maybe. Also envision the possibility, that at some point "good enough" rules the market once more, but the service with the largest impacts on media and ads simply winning out, towards even further concentration of power. Services like DAZ3D get destroyed or bought and then destroyed, we already know this version of "is not a monopoly" from the movies. Some competition is kept alive as pets and to keep regulation at bay, similar to nowadays.
    • Other jobs, other people? Partly. The scribe then can have a storyboard made with a little help by an ai - fast and efficient. Programmers who get replaced by ai, may become artists now? IDK. With all the high impact changes, ai being one, not only for the world of art, jobs undergo some re-evaluation, some new ones are there, many get less value or get removed for most, some are increased in value due to the new possibilities.
    • Jobless people everywhere? Likely - but mostly in all other jobs first. Highest impact areas first, that's not art, most is automation, just that you can automate tasks, which were considered to demand for some brain, previously. Societies will have to reform to allow for many people not having jobs all the time, because automation doesn't create more jobs in general. Rarely work is layed off to other countries, where it's cheaper. Utopia would be to increase the impact and possibilities for education in the broad, as well as for inspiring each other, and also to benefit from inspiring others. (With or without cloud services.) (There may be some cushioning effects with hauling back jobs from abroad. However large new production facilities will be built... based on the latest technology for most, i'd assume.)
    • Expectations and common sense? Hard to do common sense. Humans grow up with a high bandwidth perception, imagination (generative elements also in dreaming), and don't only learn by other people slamming images of "what can be" into their faces while shouting the relevant terms to them. We learn in many small steps, deviate based on whatever, and maybe get held back, or maybe not, high frequency feed-back from parents and people, and so much is implicit, learning from example on the one hand, but also social and social self-control happen, e.g. you're imagining or having the impulse that you could do something, but no one goes there, or your parents don't and you don't. In the end so many people end up being "different", and our expectation is, that a severely limited system, could do more than imitation and more than all different people at once and within a whimp of random training? More likely someone rolls up the market and installs some dictatorship. Should neither underestimate what these systems can do nor overestimate their abilities, the latter of which being important for societal development and not to let people assume power based on misconception.
    • Historical recordings? Maybe. Extreme cases are, humans lose abilities to do stuff, because it's "not needed". One event, and we're set back past middle age. One of my sci-fi scenarios has this for craftsmanship and production (in a remote place, detached from the center of civilization), but that's rather for the convenience, to allow me to use any assets from any period, because the blueprints come from the holo-deck for most, with the robot-ai (actual) making sense of them in a context somehow, and other robots basically printing them out. Of course there is plenty of such robots, partly due to the place having been equipped with the build-to-last self-maintaining pre-wars-tech, partly due to no one wanting to go there in the first place, plus trade being available to some extent.
    • (Memes: Rather think of "enhanced communication", maybe with generated "artwork", certainly with image modification, text display, incorporated into an easy to use interface within a social network or some application in the communication context. This is not the market-killer or artist-killer scenario, it's meant to illustrate how widespread use and acceptance could happen. Getting people used to something like  lower quality mp3 playback with all it's artifacts, can have a significant impact on the whole thing. Not directly life-threatening, but changing what can be sold to many easily. So this is more like the secondary impact of fast transitions, or a driver of a faster transition.)
    Post edited by generalgameplaying on
  • generalgameplayinggeneralgameplaying Posts: 517
    edited January 2023

    WendyLuvsCatz said:

    reply to a deleted post

    mods should delete also as made no sense

    this whole thread a dumpster fire TBH

    it was supposed to be remixing art with ai which I mostly do

    99% of my Stable Diffusion generated images use a 3D render or photograph of mine pre or post, 1% just text and no other modifications 

    Fair point. The whole discussion about impacts and odds and ends could be it's own thread. The topic is tied to it, though.

    The OP, though, hasn't posted for a while. The moment i commented first was:

    @bytescapes: "So if you render a scene and create an AI-generated image off that, and then use that image as a background for another rendered scene, and ... Where does it end?!?!???!"

    I answered something like "Do it!".

    In a way, the discussion turned the fun question, which supposedly pointed at the technical implications rather, into a societal, or just a more general one.

    Apparently both the more general discussion, as well as the one about the thread itself, have come to the same point: "Where does it end?!?!???!"

    (Oh, and on a side note: do image attachments work at the moment?)

    Post edited by generalgameplaying on
  • WendyLuvsCatzWendyLuvsCatz Posts: 38,211
    edited January 2023

    generalgameplaying said:

    WendyLuvsCatz said:

    reply to a deleted post

    mods should delete also as made no sense

    this whole thread a dumpster fire TBH

    it was supposed to be remixing art with ai which I mostly do

    99% of my Stable Diffusion generated images use a 3D render or photograph of mine pre or post, 1% just text and no other modifications 

    Fair point. The whole discussion about impacts and odds and ends could be it's own thread. The topic is tied to it, though.

    The OP, though, hasn't posted for a while. The moment i commented first was:

    @bytescapes: "So if you render a scene and create an AI-generated image off that, and then use that image as a background for another rendered scene, and ... Where does it end?!?!???!"

    I answered something like "Do it!".

    In a way, the discussion turned the fun question, which supposedly pointed at the technical implications rather, into a societal, or just a more general one.

    Apparently both the more general discussion, as well as the one about the thread itself, have come to the same point: "Where does it end?!?!???!"

    there already was an existing thread on that 

    https://www.daz3d.com/forums/discussion/441452/ai-is-going-to-be-our-biggest-game-changer/p1

    Post edited by WendyLuvsCatz on
  • generalgameplayinggeneralgameplaying Posts: 517
    edited January 2023

    WendyLuvsCatz said:

    there already was an existing thread on that 

    https://www.daz3d.com/forums/discussion/441452/ai-is-going-to-be-our-biggest-game-changer/p1

    Maybe the threads got mixed up?

    If i am right on skimming through this thread, the last post of @bytescapes actually entered the discussion themselves with something like [some] "artists not being happy", suggesting that likely some measures will be taken, mostly focusing on the "draw like [artist]" examples. This could happen too, similar to some social media and search engines, making it harder to search details about people. Artists are publishing, though, which makes them more likely targets of "mentioning". But that's only the tip of the iceberg, concerning impact on artists.

    Post edited by generalgameplaying on
  • bluejauntebluejaunte Posts: 1,902

    outrider42 said:

    To be fair, we often use AI already in some form. Denoisers, upscalers, and face reconstruction are all features based on AI software. However these are not reconstructing entire images and turning them into something else. These can do unintentionally goofy things sometimes, but they do not take a motorcycle and turn it into a goat.

    So the line for being able to copyright a work with AI will fall somewhere between that and a text to image generation. It will likely vary country to country. But I can see a lot of countries voiding text to image copyrights.

    I am not against AI art. I simply want it to be properly regulated so it doesn't get out of control. It can be fun to play with.

    It can certainly have high hardware requirements, which will no doubt be a barrier of entry. Sure, you can use various online models, but you are capped at how many free pics you can make, and their size. You also have to deal with potential censorship as seemingly innocent phrases may be banned and possibly get you banned from the service.  However doing AI on your local machine requires a ton of VRAM if you want a larger pic. If you want to train your own fork of AI you basically need 24gb of VRAM, and only a few GPUs in existance have this much. So a number of AI features are more demanding than Daz Studio Iray can ever be. As much as people in the forums complain about Daz using a lot of VRAM, these AI generators can totally obliterate some of your computers. You can use a gpu with 6 or 8gb, but you will be restricted to small images, and training is not available to you.

    So in this respect, Daz still has plenty of advantages. I can only render a maximum of 2048 by 2048 with Stable Diffusion using my 3090, this a lot larger than online AI sites allow (many are capped at 512 or 1024 pixels). But I can render significantly larger images with Iray or other render engines.

    Just use an AI image upscaler. And because I'm not confincing enough as a human, I asked gpt-3 to write me a sentence to do it. Here it is:

    "You'll be amazed at how much better your low-resolution images will look after using an AI upscaler! Not only will it save you time and effort trying to manually enhance the images yourself, but the results will be much more natural and professional-looking. Give it a try and see for yourself!"

  • outrider42outrider42 Posts: 3,679

    bluejaunte said:

    outrider42 said:

    To be fair, we often use AI already in some form. Denoisers, upscalers, and face reconstruction are all features based on AI software. However these are not reconstructing entire images and turning them into something else. These can do unintentionally goofy things sometimes, but they do not take a motorcycle and turn it into a goat.

    So the line for being able to copyright a work with AI will fall somewhere between that and a text to image generation. It will likely vary country to country. But I can see a lot of countries voiding text to image copyrights.

    I am not against AI art. I simply want it to be properly regulated so it doesn't get out of control. It can be fun to play with.

    It can certainly have high hardware requirements, which will no doubt be a barrier of entry. Sure, you can use various online models, but you are capped at how many free pics you can make, and their size. You also have to deal with potential censorship as seemingly innocent phrases may be banned and possibly get you banned from the service.  However doing AI on your local machine requires a ton of VRAM if you want a larger pic. If you want to train your own fork of AI you basically need 24gb of VRAM, and only a few GPUs in existance have this much. So a number of AI features are more demanding than Daz Studio Iray can ever be. As much as people in the forums complain about Daz using a lot of VRAM, these AI generators can totally obliterate some of your computers. You can use a gpu with 6 or 8gb, but you will be restricted to small images, and training is not available to you.

    So in this respect, Daz still has plenty of advantages. I can only render a maximum of 2048 by 2048 with Stable Diffusion using my 3090, this a lot larger than online AI sites allow (many are capped at 512 or 1024 pixels). But I can render significantly larger images with Iray or other render engines.

    Just use an AI image upscaler. And because I'm not confincing enough as a human, I asked gpt-3 to write me a sentence to do it. Here it is:

    "You'll be amazed at how much better your low-resolution images will look after using an AI upscaler! Not only will it save you time and effort trying to manually enhance the images yourself, but the results will be much more natural and professional-looking. Give it a try and see for yourself!"

    Honestly I believe GPT to be far, far more dangerous than AI art ever can be.

    Like Spiderman would say, with great power comes great responsibility. This is extremely powerful technology, and not simply a toy.

    But that is another topic.

  • kyoto kidkyoto kid Posts: 41,046
    edited January 2023

    ...last night I was at my usual neighbourhood hangout enjoying a pint of beer when fellow asked me if he could take my picture,  I replied "will it be used with AI?" after which he looked a bit disappointed and put his phone away.  

    Yeah, its happening in RL as well. and after seeing that "talking heads" avatar video above it makes me wonder more, particularly given the socio-political environment we are in.

    Post edited by kyoto kid on
  • jjmainorjjmainor Posts: 485

    Geminii23 said:

    generalgameplaying said:

    Geminii23 said:

    @valzheimer - Yeah. I've been using DAZ as a hobby for many years and recently storyboarded an entire film project with it. Many people were surprised and impressed with my storyboards and have never heard of DAZ. But I was really depressed in seeing one of my colleagues and my own brother use AI to create artwork for their own projects in record time. I love using DAZ but it made feel like I''ve wasted so much time and money all these years if people can just type some keywords into a text box and spit out great looking pre-vis artwork.

    Do they do storyboards out of nothing? 

     

    (Edited for better readability.)

    In the example of my colleague, she used Midjourney to create the artwork to accompany a 23 minute podcast. I was definitely impressed with the look and presentation which is why I started feeling depressed about how long I spent on my own storyboards for my film project. At some point will this AI make DAZ obsolete? Will comic artists be a thing of the past? Animators?  Where will it end?

     

     Following this thread since it was started, my thought was when will AI reach a point where it not only generates images in the style of a 3D render, but converts an object or objects it generates into an actual 3-D model.

  • WendyLuvsCatzWendyLuvsCatz Posts: 38,211

    I trained an embedding on my ugly face




  • generalgameplayinggeneralgameplaying Posts: 517
    edited January 2023

    jjmainor said:

     Following this thread since it was started, my thought was when will AI reach a point where it not only generates images in the style of a 3D render, but converts an object or objects it generates into an actual 3-D model.

    Entirely feasible. For some applications such exists (2d to 3d). For the arbitrary image to 3d generative ai may be needed or "interesting". Though you would need some principles for construction or training data, accordingly. Maybe the better application would be the other way round, to place your 3d models based on text and quick-render them, or a more specific tool, like create a human 3d-mesh from drawings.

    Still i could foggily envision an application where you can just do anything :), like you draw stick-figures or an image, let it be rendered differently based on text description, and either create 3d characters from it, or morph yours into them, and anyhow switch to and fro, use the image to place assets you select by hand, according to whatever your workflow then looks like (not only human characters). Such of course is a bit arbitrary and too broad for a (so i would assume) profitable endavour short-term.

     

    (For the inspiration page of things, consider generative ai, if needed at all, for applications like i draw stick figures or maybe just something, and the tool converts it into a pose for a character, with several options or "dials" for things like "realistic", "from motion", "in motion", and so on. This is a bit like the muscle morphs based on side conditions, learned from movies or sports movies. Maybe, to do this kind of thing in an efficient way, a few advancements are still needed, however it feels very much like in near-cheap reach, like "it will come more or less soon". Similar to what with fusion reactors, these kinds of things may be an interesting graph for the future, in terms of estimation vs. when it's there.)

    (Edit further: For friends of the extrapolation, based on a sarcastic format: )

    ("Oops" it puts the video inline... let's see what happens.)

    Post edited by generalgameplaying on
  • generalgameplayinggeneralgameplaying Posts: 517
    edited January 2023

    WendyLuvsCatz said:

    I trained an embedding on my ugly face

    Ugly? Interesting! Does it work with squirrels too?

    (Meaning something like... can it use squirrels instead of hats, or coats?)

    Post edited by generalgameplaying on
  • ArtiniArtini Posts: 9,457

    This video on YouTube is interesting as well:

     

  • FirstBastionFirstBastion Posts: 7,760
    edited January 2023

    Ground News - transparent bias  identified  good stuff

    Post edited by FirstBastion on
  • Interesting. Am still torn on this A.I. schtick. I've been intrigued by some of the images that have been blowing up on social media etc, but the ethics of it still bothers me, especially as an original content creator. So for example over on Artstation, since they've provided the facility to do so, I have applied the option that prevents A.I. from using my artwork to contribute to its search for content to input into its algorithm. 

    But back to your work. It certainly illustrates the limitations of the process. The least successful by far is the armoured soldier, I really hate how it has modified that image, the proprortions are all wrong, and it really hasn't added anything of significance to the original image. And with hands on show, we can see what a mess it has made of those. The weapon modification is also a little naff too. The image of the girl is much more successful, arguably giving you a more realistic image of the girl than the original. The cityscape is also quite successful, it has managed to keep the atmosphere and intent of your original image, but come up with far more complex and interesting shapes and layouts for your cityscape. I'd say that is the image that has been improved the most by the application of A.I.

  • WendyLuvsCatzWendyLuvsCatz Posts: 38,211

    generalgameplaying said:

    WendyLuvsCatz said:

    I trained an embedding on my ugly face

    Ugly? Interesting! Does it work with squirrels too?

    (Meaning something like... can it use squirrels instead of hats, or coats?)

    I tried
     

  • N-RArtsN-RArts Posts: 1,496
    edited January 2023

    Besides the ethics, it's pretty fascinating (and somewhat addictive). I have no plans to make money from it. 

    I love Midjourney, but their lack of payment options is rather off-putting when it comes to buying a subscription. So I'm settled in at Playground. Although I'm not blown away by some of the results that I've been getting. 

    Having said that, you won't be seeing my head attached/blended to the body of a squirrel anytime soon ^^' (although I would love to have that level of imagination).

     

    Post edited by N-RArts on
  • WendyLuvsCatz said:

    it's a lot easier to do stuff if your only stake in it is I have a computer, I want to have fun

    I can understand people who are using their skills to earn a living being pissed though which is why I have stated I would never use ai art commercially and don't think it should be for the reasons already discussed 

    but I also don't want people gatekeeping what I do on my own computer 

    what books I read

    what music I listen to etc

    this also applies to what programs I use and if it's not explicitly against some federal or international law it's my choice 

    if it means a lot of people on the internet will not like me because of it I won't lose as much sleep as I do about countries invading each other or people being persecuted because of their gender roles etc, it's simply not that big a deal to me in the scheme of things.

    if legislation is passed limiting what can be scraped I am perfectly fine with it and have already said I agree with that, but I also want to be part of the exciting experience pf enjoying and exploring new technology not the one in the corner hogging all the crayons and pouting 

    for me playing with ai is what I do for recreation instead of playing a game like Cyberpunk 2077 or Genshin Impact, it's not a big part of my life but a fun one

    as to sharing the results on facebook I guess those preaching never share internet memes created using popular images or share or do art involving characters from movies, comics or video games

    Having fun is one thing, having fun at no matter what cost is another. You can have fun with a gun with or without hurting people.

    I don't think anyone want to gatekeep you from creating art. People create wonderful art without arms, using their mouth or legs. I think this gatekeeping reasoning is just a fancier way to say someone being lazy and doesn't want to put any effort into what they do. Using AI is at this moment a good way to put out a lot of images with minimal effort and make money from it. Pretty much a NFT2.0 just a sleazier version that's less of a Ponzi scheme.
    Artists are pissed because their works are being used without their consent and seemingly noone is standing up for their rights. And on top of that some real pos ai wranglers come up with images that outright copy pro artist's style or pieces causing financial and reputation damage. And unlike in the world of classic or fine art, conterfeits and forgeries not only get no punishment but usually are applauded.

  • WendyLuvsCatzWendyLuvsCatz Posts: 38,211
    edited January 2023

    I am not ashamed to say I am lazy

    and not artistic

    that's why I buy DAZ assets and render them instead of drawing and painting wink

    video

    the ai embedding trained on my photos used with video footage

     

     

    Post edited by WendyLuvsCatz on
  • ArtiniArtini Posts: 9,457

    Great images and examples, @WendyLuvsCatz

    I am afraid to even start with AI remixing myself, because of my addiction to creating digital images wink

     

  • wolf359wolf359 Posts: 3,828
    edited January 2023

    It can very addicting I must confess.

    And the general public probably won’t care about the how the A I’s are “trained
    they will just enjoy the visual eye candy.

    Video Made completely with AI in ,less than 30 minutes ,including the script (chat GPT)

    Post edited by wolf359 on
  • ArtiniArtini Posts: 9,457
    edited January 2023

    Nice example. Is this voice and lip sync also created with AI?

     

    Post edited by Artini on
  • wolf359wolf359 Posts: 3,828

    Artini said:

    Nice example. Is this voice and lip sync also created with AI?


     

    Yes an AI animated a still  ( AI generated )image based on AI generated speech Audio.

  • WendyLuvsCatzWendyLuvsCatz Posts: 38,211

    they are so easily excited 

  • ArtiniArtini Posts: 9,457

    wolf359 said:

    Artini said:

    Nice example. Is this voice and lip sync also created with AI?


     

    Yes an AI animated a still  ( AI generated )image based on AI generated speech Audio.

    Great. I really like the quality of AI generated speech.

     

Sign In or Register to comment.